Kathryn Paige Harden (University of Texas) did this by looking not only at individual teens, but also at their brothers and sisters. Some of them were identical twins, but others ranged in genetic relatedness from normal siblings to half-siblings (i.e. one parent different) to cousins to adopted (unrelated).
Using this data, they were able to split up the causes of religion and also teen drinking into three factors: genetic, family environment, and external environment (the external environment is the social environment that is not shared by siblings).
Paige Harden's first finding is that only one third of the differences in religion in these American teens is down to religion. That fits in with other studies that have found that religion is mostly down to the environment in which you are raised (at least for teens, that is - genetics might be more important in later life).
She points out that this does not mean that we have genes that 'code for' religion. What it means is that, in certain cultural contexts, there are genetic factors that predispose you to adopting behaviours that fall under the label 'religion'. In a different cultural context, these same genese might act differently.
But what about religion and drinking?
Well, what she found was that religious teens (i.e. those above the middle-line of religiosity) were older on average when they had their first alcoholic drink. But the difference was tiny - religious teens postpone their drinking by only six months (from just under 14 years to just over).
However, when she looked at siblings who were raised in the same home but had different levels of religiosity, there was no difference. In other words, it's not the teenager's religious beliefs that cause the difference in drinking, but something in their environment.
Paige Harden demonstrated this more formally by showing that the family influences that are linked to religiosity are also linked to later drinking (but not the genetic factors), and that when you take this into account individual religiosity is irrelevant.
What could these family influences be? One obvious answer is that religious parents delay teen drinking. But she also found that religious mothers had no effect on the onset of teen drinking.
The most likely explanation, then, is that there is something in the wider social environment (but still shared with their family members) of teens that is linked to religion and that causes a small delay in when they choose (or are able to) take their first drink.
In other words, it's about living in an environment where teen drinking is frowned upon (which, in the highly religious USA, also tends to be religious environments).
Does this matter? Well, yes it does, because it suggests that programmes that try to make children more religious are unlikely to affect teen drinking. As the Paige Harden concludes:
These results suggest that increasing an individual adolescent’s own level of religiosity, independent of his or her family background, may not protect against early drinking initiation. Federal ‘faith-based’ initiatives, if they focus on increasing adolescent involvement in religious organizations but do not change the adolescent’s family environment, may be ineffective in preventing adolescent substance use.
Harden, K. (2010). Does religious involvement protect against early drinking? A behavior genetic approach Journal of Child Psychology and Psychiatry DOI: 10.1111/j.1469-7610.2010.02247.x
This article by Tom Rees was first published on Epiphenom. It is licensed under Creative Commons.